Motivational Interviewing Assessment Tools

MI competence is assessed by rating a recording of real or simulated practice against a structured instrument, not by asking practitioners how they think they are doing. Four instruments cover almost all of it: MITI for coded fidelity, MICA for coaching, BECCI for brief consultations, and MIA:STEP for supervisors building a review routine. They measure different things, and picking the wrong one is the usual reason an assessment tells a trainer nothing they can act on.

This guide is written for the person running the assessment rather than the person being assessed. It covers what each instrument rates, which to reach for, and how to design the sampling so the result reflects a practitioner’s actual range.

What does “MI assessment” mean?

The phrase points at two different jobs, and the confusion is worth clearing before choosing a tool.

The first is assessing the practitioner: rating someone’s MI skill against a standard. That is what this guide is about, and what MITI, MICA and BECCI do.

The second is conducting a client assessment in an MI style: taking the intake or psychosocial assessment a service already requires and running it without wrecking the working relationship. That is a clinical skill, not a measurement.

MIA:STEP contains both, which is where much of the confusion starts. Its supervisory tools rate the counsellor; its “MI Assessment Intervention” section is a protocol for the client interview, structured as three segments of roughly 30 to 45 minutes with the formal assessment sandwiched between two client-centred MI conversations. Same package, same phrase, two unrelated meanings.

The four instruments

Instrument	What it rates	Best suited to
MITI 4.2.1	Behaviour counts and four global scores from a 20-minute audio segment	Benchmarking against a published threshold; research
MICA v3.2	The flow of the conversation rather than utterance-level tallies	Coaching one practitioner over months
BECCI	Brief behaviour change consultations, on a short checklist	Short consultations; measuring change across a training
MIA:STEP	A supervisor package: skill summaries plus structured supervision	Building a supervision routine where none exists

MITI (Motivational Interviewing Treatment Integrity, currently 4.2.1) is the field’s default. A coder tallies specific behaviours across a 20-minute segment and rates four global dimensions, then compares the result against published competence and proficiency thresholds. It is the most defensible number available, and the MI fidelity and MITI coding guide covers the metrics and thresholds in detail. Its cost is what limits it: human coding is slow, and few services can afford it more than once or twice per practitioner.

MICA (Motivational Interviewing Competency Assessment, version 3.2, developed by Casey Jackson at IFIOC) was built for a different purpose. Where MITI identifies behaviours inside the conversation, MICA attends to how the conversation moves, and it was designed to feed practical coaching rather than research fidelity. Its published reliability work reports high correlation with related MITI scores, so the two are not measuring unrelated things.

BECCI (the Behaviour Change Counselling Index) came out of Cardiff, published by Lane and colleagues in Patient Education and Counseling in 2005. It rates behaviour change counselling, an MI adaptation for brief healthcare consultations, on a deliberately short checklist. Trainers use it to measure skill before and after a training, which is exactly what it was validated to do. Use it for a ten-minute consultation and it fits; use it to judge a full counselling session and it will miss most of what happened.

MIA:STEP is not a single scale. Published in 2006 through the ATTC Network under the NIDA and SAMHSA Blending Initiative, it is a public-domain package of supervisor tools, and its self-assessment skill summaries cover ten areas including MI spirit, open questions, affirmations, reflective statements, developing discrepancies and change planning. Being public domain matters more than it sounds: a service with no budget can start supervising properly this month.

Which one should you use?

Match the instrument to the decision you need to make.

Benchmarking a cohort against an external standard, or producing something an auditor will accept, means MITI. Nothing else carries the same weight.
Coaching the same practitioner over months points to MICA, because conversation flow is what improves once someone has cleared the basic counts.
Measuring what a one-day or two-day training changed, in a setting where consultations are short, points to BECCI.
Having no supervision structure at all points to MIA:STEP first, and an instrument second.

A systematic review by Hurlocker and colleagues in Clinical Psychology Review (2020) sorted the available tools into observer-rated, trainee-rated and client-rated, and found their empirical strength varies considerably, with some better suited to training and others to outcome research. The practical reading is that no instrument is the right answer to every question.

Why self-report cannot carry it

Asking practitioners to rate their own MI is the cheapest option and the least informative. Self-rated MI use correlates poorly with externally coded fidelity, which is the reason MI training invests so heavily in recording and review, and the fidelity guide sets out the evidence.

Self-assessment still earns its place, just not as the measurement. It is useful for directing attention: a practitioner who has worked through the MIA:STEP skill summaries knows what a complex reflection is supposed to do, and arrives at supervision able to discuss their own session rather than waiting to be told. Use it to prepare people for assessment. Do not use it as the assessment.

Designing the assessment

The instrument is the easy part. Most assessments that fail do so in the sampling, and none of the manuals will warn you.

Do not let practitioners choose the recording

This is the largest single source of error. A practitioner picking which session to submit picks one that went well, so the result describes their ceiling and says nothing about their floor. Ask for a defined window instead, several consecutive sessions across a fortnight, and select from it yourself. The point is not to catch anyone out. It is that a cohort’s weakest sessions are where the training need actually is.

One session is a data point, not a picture

MITI codes a 20-minute segment. Performance varies across a session and varies far more across clients, so a single coded sample tells you about that sample. Two or three across different clients is the minimum before a trainer should describe a practitioner’s competence at all, and before anyone announces that a cohort has improved.

Calibrate the raters before rating anyone

Two supervisors watching the same recording routinely disagree, particularly on the global ratings, where partnership and empathy leave real room for judgement. Behaviour counts hold up better than globals. Have every rater code the same session independently, compare, and argue it out before any of them rate a live cohort. Repeat it periodically, because rater drift is quiet and cumulative.

Decide in advance what the result is for

Formative and summative assessment pull in opposite directions. If a score determines whether someone passes a course, practitioners will submit their safest work and the assessment stops measuring practice. If it feeds coaching and carries no consequence, people submit the session that troubled them, which is the useful one. Trainers who want both usually need two separate exercises, not one score doing double duty.

Turning a score into practice

A coded score identifies where someone is. It does not move them, and the gap between the two is where most MI training quietly fails: a workshop, one coded session, a report, then nothing for a year while the skill decays.

What closes that gap is repetition with feedback between assessments, which is the part traditional supervision cannot supply at volume. A supervisor who codes one session a quarter is doing something valuable and something rare. They are also giving the practitioner four data points a year and no reps in between, and MI skills decay without practice.

The MI Practice Lab is built for that middle space. Practitioners hold voice conversations with realistic clients and get behaviour-count feedback immediately, with the client’s own words quoted alongside each judgement so they can check the reasoning rather than trust a number. It is formative feedback, not a credential, and it does not replace a coded assessment. It fills the eleven weeks between one and the next. For trainers running a cohort, the trainer overview covers how that fits alongside supervision.

Frequently Asked Questions

How is competence in motivational interviewing assessed?

By rating a recording of a real or simulated session against a structured instrument. The most widely used is MITI 4.2.1, where a trained coder tallies specific behaviours across a 20-minute audio segment and rates four global dimensions, then compares the result against published competence and proficiency thresholds. Self-report is not a substitute, because self-rated MI use correlates poorly with externally coded fidelity.

Is there a motivational interviewing self assessment?

Yes, and MIA:STEP is the best-known free one: its self-assessment skill summaries cover ten areas including MI spirit, open questions, affirmations, reflective statements and change planning, with descriptions of higher and lower skill levels so a practitioner can locate themselves. Treat it as preparation rather than measurement. It is genuinely useful for directing attention before supervision, and it should not be the evidence a trainer relies on, because practitioners consistently over-rate their own MI.

What is the difference between MITI and MICA?

MITI identifies specific behaviours and utterances inside the conversation and produces counts and global scores against published thresholds, which makes it the stronger choice when you need a defensible benchmark. MICA attends more to how the conversation flows and was designed to feed practical coaching across service settings rather than research fidelity. Reliability work on MICA reports high correlation with related MITI scores, so they broadly agree about who is skilled; they differ in what they hand back to the practitioner.

How many sessions does a fair assessment need?

More than one. A MITI-coded segment covers 20 minutes, and performance varies both across a single session and much more across different clients, so one sample describes that sample rather than the practitioner. Two or three recordings with different clients is a reasonable floor before describing someone’s competence, and letting the practitioner choose which sessions to submit undermines the whole exercise because people select their best work.

Is MIA:STEP still worth using?

Yes, particularly where there is no budget and no existing supervision structure. It was published in 2006 through the ATTC Network under the NIDA and SAMHSA Blending Initiative and placed in the public domain, so a service can adopt and adapt it without licensing anything. Its age shows in places, and it predates MITI 4, so pair its supervision structure with a current instrument rather than relying on it alone for scoring.

Does an MI assessment certify a practitioner?

Not in any formal sense in most jurisdictions, because there is no central MI certification body. The Motivational Interviewing Network of Trainers does not require coded scores for membership, and most training programmes use MITI or MICA formatively rather than as a pass or fail gate. Some employers and funders set their own internal thresholds, which is a local requirement rather than a recognised credential.

Want to give a cohort reps between coded sessions? The MI Practice Lab gives practitioners behaviour-count feedback on every conversation, with the client’s words quoted next to each judgement so they can check the reasoning. Start a free trial: 5 minutes, no card required.